Creating “Quicksand Themed Video Games” With ChatGPT-6 Astra?

[Quicksand-Lover-7]

So a few days ago. OpenAI released ChatGPT-6 Astra. And it looks like it can do a lot of things. And apparently it can also create video games and 3D models. And I don’t know how powerful or complex it actually is since I haven’t used it myself. But if that’s actually true that you can ask it to create a video game and it can do that. Then imagine all the simple games that people will start creating with the AI.

Anyways, so I wonder if you could just turn on Unreal Engine and ask it.

“Hey ChatGPT-6 Astra, go to Unreal Engine and create a 3D platformer video game where you play as a Jungle Woman in a jungle. But there are lots of quicksand hazards that the player has to avoid. And if the player steps in quicksand. They have to struggle to escape before they sink all the way down and lose.”

So if you asked it that, I wonder if it would actually be able to do it. That would be really interesting. Or a Tomb Raider like game, where Lara Croft can sink into quicksand.

[Theo]

We Needs IT!.jpg

[Quicksand-Lover-7]

I found some interesting videos of people creating video games using ChatGPT-6 Astra. I thought they seemed interesting. It looks like it’ll make game creation a lot faster and more accessible for people who don’t have much experience with programming. I know that there are a few people in the quicksand community who make quicksand themed games and I wonder if some are going to try using this to make games. Although it probably still takes more time to ask an AI to create a game than to create a video or an image. Maybe later on in the future, I’ll try to create a quicksand themed video game using it, and see how well it turns out.

Here are the 3 Links:
https://m.youtube.com/watch?v=GuO_Eo34C ... 3D%3D&ra=m

https://m.youtube.com/watch?v=HVrwoywwv ... 3D%3D&ra=m

https://m.youtube.com/watch?v=HOKOrTtPz ... 3D%3D&ra=m

And some screenshots and last quicksand images.

[Aiko]

I found some interesting videos of people creating video games using ChatGPT-6 Astra. I thought they seemed interesting. It looks like it’ll make game creation a lot faster and more accessible for people who don’t have much experience with programming. I know that there are a few people in the quicksand community who make quicksand themed games and I wonder if some are going to try using this to make games. Although it probably still takes more time to ask an AI to create a game than to create a video or an image. Maybe later on in the future, I’ll try to create a quicksand themed video game using it, and see how well it turns out.

Here are the 3 Links:
https://m.youtube.com/watch?v=GuO_Eo34C ... 3D%3D&ra=m

https://m.youtube.com/watch?v=HVrwoywwv ... 3D%3D&ra=m

https://m.youtube.com/watch?v=HOKOrTtPz ... 3D%3D&ra=m

And some screenshots and last quicksand images.

I think the first video already shows the issues with AI coding quite well. You spend 250 USD on something that looks interesting but nobody would actually want to play. Of course you might say that this is only the beginning of the project and you can get in all the changes you want with more prompts. What you should actually expect to happen is that the AI will try to implement those changes with varying degrees of success and it will start breaking other things that were already OK. Unless you really understand the code and what the AI is doing, you will have a very hard time addressing underlying issues. So if you are ready to spend thousands of dollars on an experiment that might not turn out that way you are hoping at all (or you are actually prepared to learn about programming to take over and solve things the AI cannot), then go ahead. If I still have not discouraged you from trying, at least keep things very, very simple (only 2D, few gameplay mechanics).

[LobsterApple]

To add to what Aiko is saying, I think making an AI mod of a pre-existing game is probably cheaper and better than just getting the AI to do all the work from scratch, since at least everything is built on top of an already functioning base and the AI is not sophisticated enough yet to build everything at a feasible scale. And if things don't work out, at least it's less money down the drain.

Since you were interested in making a Tomb Raider style game, I would start by asking AI to code a mod that adds quicksand mechanics to one of the pre-existing Tomb Raider games. Otherwise, Aiko is right in that doing anything from scratch with AI at this point in time would be far too complex to do anything more than 2D games properly. I would like to see how these experiments turn out, but I do think you need to use AI intelligently rather than just everywhere and anywhere. Have Fun.

EDIT: Found 2 videos that were using the AI process I was talking about, though with GTA San Andreas in this case:
https://youtu.be/aYeXEVkz1Ig?si=k8WyuckGURAAOcoQ
https://youtu.be/isjBe4khZZc?si=fMb3WCdFoMiplXtA

[Aiko]

To add to what Aiko is saying, I think making an AI mod of a pre-existing game is probably cheaper and better than just getting the AI to do all the work from scratch, since at least everything is built on top of an already functioning base and the AI is not sophisticated enough yet to build everything at a feasible scale. And if things don't work out, at least it's less money down the drain.

Since you were interested in making a Tomb Raider style game, I would start by asking AI to code a mod that adds quicksand mechanics to one of the pre-existing Tomb Raider games. Otherwise, Aiko is right in that doing anything from scratch with AI at this point in time would be far too complex to do anything more than 2D games properly. I would like to see how these experiments turn out, but I do think you need to use AI intelligently rather than just everywhere and anywhere. Have Fun.

EDIT: Found 2 videos that were using the AI process I was talking about, though with GTA San Andreas in this case:
https://youtu.be/aYeXEVkz1Ig?si=k8WyuckGURAAOcoQ
https://youtu.be/isjBe4khZZc?si=fMb3WCdFoMiplXtA

You are partially right. Modding an existing game can be a lot cheaper/less work than creating one from scratch (does not matter whether you do it yourself or let an AI do it for you). However, that is only true if the game is well-suited for modding, and for Tomb Raider that does not seem to be the case. The few community-made modding tools I found can help you replace some assets like textures or 3D models inside the game, but that sort of modding does not get you anywhere, if you want to introduce new game mechanics. Eg. you might be able to make the water surface look like quicksand, but Lara will still be swimming/diving in it like before. Given that TR uses a closed-source proprietary engine, modding game mechanics is a huge challenge if possible at all. You could try and look for a game that has well-documented modding support and allows that sort of thing. Or you could take an open source game and tell the AI to add quicksand to it. If you have the source code, you/the AI can change anything. I did a quick search for suitable projects and found this 2D platformer that might work (https://johngabrieluk.itch.io/hijinx-enhanced-edition). The project seems small enough so that an AI has a chance to make sense of the codebase and add features to it. That seems a whole lot more realtistic than asking it to mod/build TR with quicksand.